JP6868057B2 - Reading system, reading method, program, storage medium, and mobile - Google Patents

Description

Embodiments of the present invention relate to reading systems, reading methods, programs, storage media, and mobiles.

There is a system that reads the characters (for example, numbers) displayed on the segment display. It is desired to improve the reliability of the reading system.

Japanese Unexamined Patent Publication No. 2001-291554

An object to be solved by the present invention is to provide a reading system, a reading method, a program, a storage medium, and a mobile body capable of improving reliability.

The reading system according to the embodiment includes a reading unit and a calculation unit. The reading unit reads the characters from a character image of the characters displayed by the segment display including the plurality of segments. The calculation unit executes the first process of calculating the first score, and calculates the reading accuracy using the first score . In the first process, the calculation unit sets a plurality of determination regions corresponding to the plurality of segments in the character image, and includes each of the plurality of determination regions in the thin-lined pixels indicating the characters. The first score is calculated based on the number of lines to be drawn and the angle of the lines overlapping the determination area.

It is a block diagram which shows the structure of the reading system which concerns on embodiment. It is a figure which illustrates the process by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a table exemplifying the correspondence between the combination of judgment results and numbers. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a figure for demonstrating the processing by the reading system which concerns on embodiment. It is a flowchart which illustrates the process by the reading system which concerns on embodiment. It is a graph which shows the reading result and accuracy in an Example. It is a schematic diagram which illustrates the setting of a segment display and a determination area. It is a block diagram which shows the structure of the reading system which concerns on embodiment. It is a block diagram which shows the structure of another reading system which concerns on embodiment. It is a schematic diagram explaining the operation of another reading system which concerns on embodiment. It is a flowchart which illustrates the process by the reading system which concerns on embodiment. It is a flowchart which illustrates the process by the reading system which concerns on embodiment.

Hereinafter, each embodiment of the present invention will be described with reference to the drawings.
In the present specification and each figure, the same elements as those already described are designated by the same reference numerals, and detailed description thereof will be omitted as appropriate.

FIG. 1 is a block diagram showing a configuration of a reading system according to an embodiment.
The reading system according to the embodiment is used to read the characters displayed on the segment display from the image captured by the segment display. In the segment display, one character is displayed by a plurality of segments. Characters are displayed by turning on at least a part of multiple segments and turning off the remaining segments. The reading system according to the embodiment reads this character from the image. A plurality of characters may be displayed on the segment display. In this case, each of the plurality of characters is displayed by lighting at least a part of the plurality of segments.

The number of segments included in the segment display to be read is arbitrary. For example, the segment display to be read is a so-called 7-segment display in which one character is displayed by seven segments. The 7-segment display displays letters (numbers) indicating numbers. A 14-segment display or a 16-segment display displaying an alphabet may be a reading target. Here, a case where the numbers displayed on the 7-segment display are read by the reading system according to the embodiment will be mainly described.

As shown in FIG. 1, the reading system 1 according to the embodiment includes a processing device 10. The processing device 10 includes a reception unit 11, a preprocessing unit 12, an extraction unit 13, a thinning unit 14, a reading unit 15, a calculation unit 16, and an output unit 17.

2 (a) to 2 (d) are diagrams illustrating processing by the reading system according to the embodiment.
An image captured by the segment display is input to the processing device 10. The reception unit 11 receives the input image. For example, an external imaging device captures a segment display and generates an image. The image pickup apparatus transmits an image to the processing apparatus 10. Alternatively, the image may be transmitted from the external storage device to the processing device 10. The reception unit 11 receives the image transmitted to the processing device 10. The image may show something other than the segment display. Here, the image received by the reception unit 11 is referred to as an input image.

FIG. 2A is an example of the input image A transmitted to the processing device 10. In the example of FIG. 2A, a plurality of numbers are displayed on the segment display. Each number is represented by multiple segments.

The preprocessing unit 12 applies preprocessing to the input image before reading the numbers displayed on the segment display. Preprocessing can improve the accuracy of reading numbers. For example, the pretreatment unit 12 includes a binarization unit 12a, an expansion processing unit 12b, and a shrinkage processing unit 12c.

The binarization unit 12a binarizes the input image. The binarized input image is represented by two different colors (first color and second color). Here, the case where the first color is white and the second color is black will be described.

For example, the background of the segment display is a dark color. The lit segment is brighter than the background. Therefore, in the binarized input image, the lit segment is represented in white. The off-lit segments and background are shown in black. That is, in the binarized input image, numbers are indicated by white pixels.

Alternatively, in a segment display using a liquid crystal display, the characters are represented by a color darker than the background. In this case, in the binarized input image, the color of each pixel is inverted. By inverting, the numbers are indicated by the white pixels, as in the case described above. Hereinafter, a case where the lit segment is represented in white and the lit segment and the background are represented in black will be described.

For example, the input image is digitized as an RGB color model shown using three primary colors of red (Red), green (Green), and blue (Blue). In the binarization process, first, this input image is converted into data shown in an HSV color space composed of three components of hue, saturation, and luminance. Next, a histogram analysis of the data in the HSV color space is performed. Subsequently, the threshold value is calculated based on the histogram of each pixel. Based on this threshold value and the histogram of each pixel, each pixel is binarized into white and black.

The expansion processing unit 12b expands the white pixels in the binary image. For example, the expansion processing unit 12b changes the pixel adjacent to the white pixel to white. If the adjacent pixels are originally white, they are not changed. For example, due to the influence of noise or the like, black pixels may be scattered in a portion where white pixels are gathered. By the expansion process, the scattered black pixels are changed to white pixels. As a result, the influence of noise and the like can be reduced, and the reading accuracy of numbers can be improved. Further, by the expansion process, it is possible to connect a set of adjacent white pixels. By connecting a set of adjacent white pixels to each other, a set of one pixel is generated corresponding to one number. This facilitates the extraction of the character image described later.

The shrinkage processing unit 12c shrinks the white pixel in the binary image. For example, the shrinkage processing unit 12c changes the pixel adjacent to the black pixel to black. If the adjacent pixel is originally black, it is not changed. The shrinkage process reduces the number of expanded white pixels.

FIG. 2B is an example of the input image A that has been preprocessed (binarized, expanded, and contracted). As shown in FIG. 2B, these preprocessings generate a set of white pixels corresponding to each number.

The pretreatment unit 12 may execute the above-mentioned expansion treatment and contraction treatment a plurality of times. For example, the pretreatment unit 12 executes the expansion process twice or more, and then executes the contraction process twice or more. The pretreatment unit 12 may execute the expansion treatment twice or more after executing the contraction treatment twice or more. For example, the number of times the expansion process is executed and the number of times the contraction process is executed are set to be the same. Alternatively, the number of times the expansion process is executed may be different from the number of times the contraction process is executed. For example, the number of executions of the expansion process is set to be larger than the number of executions of the contraction process.

The pretreatment unit 12 may execute a treatment set including one expansion treatment and one contraction treatment a plurality of times. In one set of treatments, one of the expansion and contraction treatments is performed and then the other. The order of expansion and contraction treatments in one treatment set may be different from the order of expansion and contraction treatments in another treatment set.

The preprocessing unit 12 may further execute another process. For example, when the input image includes an object other than the segment display, the preprocessing unit 12 may cut out a portion where the segment display is captured from the input image. If the input image is distorted, the preprocessing unit 12 may correct the distortion.

The preprocessing unit 12 outputs the processed input image to the extraction unit 13. The extraction unit 13 extracts a character image from the input image. The character image is a part of the input image, and one number displayed by the segment display is copied.

For example, the extraction unit 13 includes a labeling processing unit 13a and a cutting unit 13b. The binarized input image is output from the preprocessing unit 12 to the extraction unit 13. The labeling processing unit 13a assigns a label (value) to the set of white pixels. The “set of white pixels” refers to a portion in which white pixels are adjacent to each other and form a single white mass. In a set of white pixels, a white pixel is adjacent to at least one white pixel. One set of white pixels corresponds to one number displayed on the segment display. When there are a plurality of sets of white pixels, the labeling processing unit 13a assigns a label to each set of white pixels.

The cutout portion 13b cuts out a portion including the labeled set from the input image. The cut out part becomes a character image. When there are a plurality of labeled sets, the cutout unit 13b cuts out a plurality of character images. Each of the plurality of character images contains a plurality of labeled sets. For example, the shape of the character image is a quadrangle. The character image includes a plurality of pixels arranged in a first direction and a second direction intersecting the first direction. For example, the character image is cut out so that its size becomes a preset value. The size of the character image to be cut out corresponds to the size of the mask described later. FIG. 2C shows the input image A and the character images B1 to B4.

The extraction unit 13 outputs the extracted (cut out) character image to the thinning unit 14. The thinning unit 14 performs thinning processing on the character image. That is, the thinning unit 14 processes the set of white pixels included in the character image so that the line width is one pixel. FIG. 2D shows the result of thinning the character images B1 to B4 represented by FIG. 2C.

The thinning unit 14 may thin the binary image output from the preprocessing unit 12. For example, when the character image is cut out, the cutout portion 13b saves the cut out position. The thinning unit 14 outputs the thinned binary image to the extraction unit 13. The cutout portion 13b cuts out a part of the thinned binary image at the saved cutout position. As a result, the thinned character image shown in FIG. 2D can be obtained.

The reading unit 15 reads characters from, for example, a thinned character image. The reading unit 15 may read characters from the character image before thinning. The calculation unit 16 calculates the score using the extracted character image. Specifically, the calculation unit 16 executes the following first to seventh processes. Further, the calculation unit 16 calculates the reading accuracy by using the first score to the seventh score obtained by executing the first to seventh processes.

In the following, an example will be described in which each score and accuracy are calculated higher as the character image is better and the number is more likely to be read correctly. The higher the score and accuracy, the more reliable the reading. A good character image means, for example, a state in which the character image does not contain anything other than characters, the resolution is sufficiently high, and the distortion is small. Not limited to this example, the better the character image, the lower each score and accuracy may be calculated. In this case, the lower the score and accuracy, the higher the reliability of the reading.

3 (a), 3 (b), 4 (a) to 4 (f), 5 (a) to 5 (c), 6 (a) to 6 (e), 7 (A), FIGS. 7 (b), 8 (a) to 8 (e), 10 (a) to 10 (c), 11 (a), and 11 (b) are embodiments. It is a figure for demonstrating the processing by the reading system which concerns on.
FIG. 9 is a table illustrating the correspondence between the combination of determination results and the numbers.

In the first process, the calculation unit 16 sets a plurality of determination areas in the thinned character image. The number of determination areas to be set may be equal to or greater than the number of segments used for displaying one character. For example, the number of determination areas set is the same as the number of segments used to display one number. For example, when reading a number displayed on a 7-segment display, seven determination areas are set in one character image. The position of each determination area is determined based on the size of the extracted character image. For example, the storage device 20 stores information regarding the set position of the determination area. The calculation unit 16 determines the position and size of the determination area based on the information stored in the storage device 20 and the size of the character image. Alternatively, when the size of the character image to be cut out is predetermined, the position and size of each determination area may be predetermined.

Here, an example in which the position and size of the determination area are determined based on the size of the character image will be described. FIG. 3A is a schematic diagram for explaining a method of setting a determination area. The character image B shown in FIG. 3A is composed of a plurality of pixels P. Judgment areas DR1 to DR7 are set in the character image B. The determination areas DR1 and DR2 are arranged along the first direction D1 (horizontal direction). The determination regions DR3 and DR4 are arranged along the first direction D1. The determination areas DR1 and DR2 are separated from the determination areas DR3 and DR4 in the second direction D2 (vertical direction). The determination areas DR5 to DR7 are arranged along the second direction D2.

As shown in the determination regions DR5 to DR7 shown in FIG. 3A, the arrangement direction of some of the determination regions may not be parallel to the second direction D2 but may be inclined with respect to the second direction D2. For example, the angle between the arrangement direction of a part of the determination region and the second direction D2 is set to be larger than 0 degrees and 20 degrees or less. Similarly, the arrangement direction of another part of the determination region may be tilted with respect to the first direction D1. For example, the angle between the arrangement direction of another part of the determination region and the first direction D1 is set to be larger than 0 degrees and 20 degrees or less. Such an arrangement is also included when a plurality of determination regions are arranged along the first direction D1 or the second direction D2.

Let L1 be the length (number of pixels) of the character image B shown in FIG. 3A in the first direction D1. Let L2 be the length of the character image B in the second direction D2. Further, the character image B has a first side S1 to a fourth side S4. The first side S1 and the second side S2 are parallel to the first direction D1. The third side S3 and the fourth side S4 are parallel to the second direction D2.

For example, the determination areas DR1 and DR2 are set with reference to a position L2 / 4 away from the first side S1. For example, the determination areas DR3 and DR4 are set with reference to a position L2 / 4 away from the second side S2. For example, the determination areas DR5 to DR7 are set with reference to an intermediate position between the third side S3 and the fourth side S4. For example, the determination areas DR5 and DR7 are set at positions deviated from the intermediate position in the first direction D1. For example, the length of each of the determination regions DR5 to DR7 in the second direction D2 is set to L2 / 3. Information about these reference positions is stored in, for example, a storage device 20. Further, the determination area is set so as not to overlap with another determination area. Each determination area includes a pixel P at a reference position and pixels P around the pixel P. The size of each determination area is set according to, for example, the size of the character image B.

The reference position for setting the determination area with respect to the size of the character image and the size of the determination area with respect to the size of the character image can be appropriately changed according to the segment display to be read, the characteristics of the input image, and the like. The positions of the plurality of determination regions may be determined based on the size of the character image (length in the first direction and length in the second direction).

The calculation unit 16 determines the state of the thinned pixels in each determination area. As the pixel state, the number of lines overlapping the determination area is used. A line is a part of a plurality of pixels indicating a number, and refers to a series of pixels. For example, in the examples shown in FIGS. 3A and 3B, if the character image is clear, the number of lines detected in each determination area is at most one. If two or more lines are detected in one determination area, the character image may be unclear, the characters may be distorted, or the characters may not be correctly copied.

The calculation unit 16 calculates the first score based on the state of the pixels in each determination area. For example, the calculation unit 16 calculates the first score lower as more lines are detected in one determination area. Further, the calculation unit 16 calculates the first score lower as the number of determination regions in which two or more lines are detected increases. The calculation unit 16 may lower the first score when the line is interrupted in the determination area. The calculation unit 16 may lower the first score when there are other discrete lines near one line.

The calculation unit 16 may use the angle of the line as the pixel state. For example, in the examples shown in FIGS. 3A and 3B, if the character image is clear, a line along the second direction D2 is detected in the determination areas DR1 to DR4. In the determination areas DR5 to DR7, a line along the first direction D1 is detected. The calculation unit 16 calculates the first score lower as the line detected in each determination region is tilted with respect to the preset direction.

Alternatively, the calculation unit 16 may calculate the first score based on both the number of lines and the angle of the lines in each determination area. This makes it possible to increase the reliability of the accuracy calculated using the first score.

4 (a), 4 (c), and 5 (a) are examples of character images. FIG. 4B is a part of the character image of FIG. 4A. FIG. 4B is an enlarged representation of the determination area DR5 set for the character image of FIG. 4A. 4 (d) and 4 (e) are a part of the character image of FIG. 4 (c). FIG. 4D shows an enlarged determination area DR5 set for the character image of FIG. 4C. FIG. 4E is an enlarged representation of the determination area DR6 set for the character image of FIG. 4C. 5 (b) and 5 (c) are a part of the character image of FIG. 5 (a). FIG. 5B is an enlarged representation of the determination area DR6 set for the character image of FIG. 5A. FIG. 5C is an enlarged representation of the determination area DR7 set for the character image of FIG. 5A.

The calculation unit 16 searches for a line extending in a predetermined direction in each determination region. Specifically, the calculation unit 16 searches for a line extending in the second direction D2 in the determination areas DR1 to DR4 extending in the first direction D1. The calculation unit 16 searches for a line extending in the first direction D1 in the determination areas DR5 to DR7 extending in the second direction D2.

In the example of FIG. 4B, the line Li1 in which the pixels are connected is along the first direction D1. Further, the number of lines existing in the determination area DR5 is one. Therefore, the first score is calculated high with respect to the determination area DR5.

In the example of FIG. 4D, a plurality of lines Li2 and Li3 along the first direction D1 exist in the determination region DR5. Therefore, the first score is calculated low with respect to the determination area DR5.
In the example of FIG. 4 (e), a plurality of lines Li4 and Li5 are present in the determination region DR6. The position of the end of the line Li4 in the first direction D1 coincides with the position of the end of the line Li5 in the first direction D1. When the distance between the lines Li4 and Li5 in the second direction D2 is short, it is considered that these lines should be represented by a single line. For example, when the distance between the lines Li4 and Li5 in the second direction D2 is equal to or less than a preset value, the calculation unit 16 determines that the lines to be displayed by one line are discrete, and lowers the first score. calculate.

In the example of FIG. 5B, the line Li6 is interrupted in the determination region. In the example of FIG. 5C, the angle of the line Li7 is inclined with respect to the first direction D1. In the example shown in FIG. 5B or FIG. 5C, the first score is calculated low for the determination areas DR6 and DR7, respectively.

In the second process, the calculation unit 16 refers to the storage device 20 and calculates the precision ratio using one character image and the mask. The precision rate indicates how clearly characters based on multiple segments appear in the character image.

First, the calculation unit 16 calculates the logical product of the character image and the mask. The mask corresponds to a state in which all of a plurality of segments for displaying one number are lit. In the processed image obtained by the logical product, only the region represented in white in the mask and represented in white in the character image is represented in white.

The calculation unit 16 calculates the matching rate of the character image with respect to the processed image as the matching rate. That is, the ratio (a1 / a2) of the area a1 of the region represented by white in the processed image to the area a2 of the region represented by white in the character image is calculated as the precision ratio. The area in each image corresponds to the number of pixels represented in white. Instead of this calculation, the calculation unit 16 may calculate the ratio of the area of the area represented by black in the character image to the area of the area represented by black in the processed image as the precision ratio.

6 (a) and 6 (d) are examples of character images. 6 (b) and 6 (e) are examples of masks. 6 (c) and 6 (f) are images showing the result of the logical product.
As an example, the processed image shown in FIG. 6 (c) can be obtained by the logical product of the character image shown in FIG. 6 (a) and the mask shown in FIG. 6 (b). The calculation unit 16 calculates the precision ratio using the image shown in FIG. 6 (a) and the image shown in FIG. 6 (c). As another example, the processed image shown in FIG. 6 (f) can be obtained by the logical product of the character image shown in FIG. 6 (d) and the mask shown in FIG. 6 (e). The calculation unit 16 calculates the precision ratio using the image shown in FIG. 6 (d) and the image shown in FIG. 6 (f).

The change from the image of FIG. 6 (a) to the image of FIG. 6 (c) is smaller than the change from the image of FIG. 6 (d) to the image of FIG. 6 (f). Therefore, the conformity rate of the image of FIG. 6A is calculated to be higher than that of the image of FIG. 6D. For example, the calculation unit 16 outputs the calculated precision rate as the second score. That is, the higher the precision rate, the higher the second score is output. The calculation unit 16 may output another value calculated based on the precision rate as the second score.

In the third process, the calculation unit 16 calculates the ratio of the lengths of the character images. The calculation unit 16 calculates the third score higher as the ratio is closer to the preset value. The value corresponds to the ratio of the lengths of the set of segments to display one number.

For example, with respect to the character image shown in FIG. 7A, the calculation unit 16 calculates the ratio of the length L3 in the first direction D1 and the length L4 in the second direction D2. The character image of FIG. 7A is clear and the distortion of the copied characters is small. Therefore, the difference between the ratio of the length L3 and the length L4 and the preset value is small. The calculation unit 16 calculates a high third score for the character image shown in FIG. 7A.

On the other hand, in the character image shown in FIG. 7B, the length of the first direction D1 is long due to the influence of light reflection on the segment display. Therefore, the difference between the ratio of the length L5 and the length L6 and the preset value is large. The calculation unit 16 calculates the third score low for the character image shown in FIG. 7B. In addition to the reflection of light, the ratio of the length in the first direction D1 to the length in the second direction D2 changes from the preset value due to the influence of deposits on the segment display, noise in the image, and the like. sell.

In the fourth process, the calculation unit 16 sets a plurality of determination areas in one character image as in the first process. The method of setting the determination area is the same as that of the first process. The calculation unit 16 detects the presence or absence of a part of the character in each determination area. As a specific example, the calculation unit 16 detects the number of pixels indicating characters in each determination area. When the lit segment is represented in white, the calculation unit 16 detects the number of white pixels in each determination region. The calculation unit 16 compares the detected number of pixels with a preset threshold value, and determines whether or not a part of the numbers exists in each determination area. Specifically, if the number of detected pixels is equal to or greater than the threshold value, the calculation unit 16 determines that a part of the numbers exists in the determination area.

For example, when setting a plurality of determination areas for a thinned character image, the threshold value is set to 1. When setting a plurality of determination regions for a character image that is not thinned, the threshold value may be a value larger than 1.

The calculation unit 16 indicates, for example, the determination result in each determination area by “0” or “1”. “0” indicates that a part of the number does not exist in the determination area. “1” indicates that a part of the number exists in the determination area. In the example of FIG. 3B, it is determined that a part of the numbers is present in the determination areas DR2, DR3, and DR5 to DR7. This result is represented as "0110111" using, for example, 0 and 1.

8 (a) to 8 (e) show a character image in which other numbers are copied, and an example of setting a determination area in each character image. Similar to FIG. 3B, a plurality of determination areas are set for the character images shown in FIGS. 8A to 8E, and the presence or absence of a part of the numbers is determined in each determination area. To.

The calculation unit 16 refers to the storage device 20. The storage device 20 stores the type of each number that can be displayed on the segment display. The type is represented by using 0 and 1 as in the combination of the determination results by the calculation unit 16. FIG. 9 illustrates a type stored in the storage device 20. The calculation unit 16 searches for a type that matches the combination of determination results.

For example, in the example of FIG. 3B, the combination of determination results is represented as "0110111". The calculation unit 16 searches for a type that matches this combination. As a result of the search, the combination of the determination results by the calculation unit 16 matches the type corresponding to the number "2". When there is a matching type, the calculation unit 16 calculates the fourth score higher. When there is no matching type, the calculation unit 16 calculates the fourth score low.

In the fifth process, the calculation unit 16 calculates the fifth score based on the variation in the positions of the plurality of characters and the arrangement directions of the plurality of characters.

First, the calculation unit 16 calculates the position of the center of gravity of each of the plurality of character images in the input image. The calculation unit 16 generates a straight line based on a plurality of positions of the center of gravity. The straight line is, for example, a regression line at a plurality of center of gravity positions. The calculation unit 16 uses a regression line to obtain variations in the positions of a plurality of centers of gravity. The variability is represented, for example, by the standard deviation or the mean square error. The calculation unit 16 calculates the fifth score lower as the variation becomes larger.

Further, the calculation unit 16 calculates the difference between the arrangement direction of the plurality of characters and the preset reference direction. The calculation unit 16 calculates the fifth score lower as the difference increases. The character arrangement direction is represented by, for example, a regression line. The calculation unit 16 uses the angle between the line along the reference direction and the regression line as the difference between the arrangement direction and the reference direction. The calculation unit 16 may compare the angle with a preset threshold value. The calculation unit 16 calculates the fifth score low when the angle exceeds the threshold value.

For example, the calculation unit 16 calculates the position of the center of gravity for each of the plurality of character images B5 to B7 shown in FIG. 10A. As a result, for example, the positions C5 to C7 of the center of gravity are calculated. The calculation unit 16 generates a straight line CL based on the positions C5 to C7 of the center of gravity. For example, the reference direction is set parallel to the first direction D1. In the example shown in FIG. 10A, the variation of the center of gravity positions C5 to C7 is small. Further, the straight line CL is along the first direction D1. Therefore, the fifth score is calculated high for the plurality of character images shown in FIG. 10A.

In the example shown in FIG. 10B, the straight line CL follows the first direction D1, but the center of gravity positions C5 to C7 vary widely. Therefore, the fifth score based on the plurality of character images shown in FIG. 10B is calculated to be lower than the fifth score based on the plurality of character images shown in FIG. 10A. Further, in the example shown in FIG. 10C, the straight line CL is tilted with respect to the first direction D1. Therefore, the fifth score based on the plurality of character images shown in FIG. 10B and the fifth score based on the plurality of character images shown in FIG. 10C are a plurality of scores shown in FIG. 10A. It is calculated to be lower than the fifth score based on the character image of.

In the sixth process, the calculation unit 16 compares the number of characters copied in the image with the preset minimum number of characters and maximum number of characters. For example, the calculation unit 16 compares the number of character images (the number of labeled sets) with the minimum number of characters and the maximum number of characters. The calculation unit 16 calculates the sixth score higher when the number of characters is equal to or greater than the minimum number of characters and equal to or less than the maximum number of characters. The calculation unit 16 calculates the sixth score low when the number of characters is less than the minimum number of characters or exceeds the maximum number of characters.

In the seventh process, the calculation unit 16 determines whether the decimal point is recognized in the input image. In the storage device 20, whether the segment display displayed in the input image includes a decimal point is registered in advance. The calculation unit 16 compares the recognition result with the information registered in the storage device 20. When the recognition result matches the information registered in the storage device 20, the calculation unit 16 calculates the seventh score high. When the recognition result does not match the information registered in the storage device 20, the calculation unit 16 calculates the seventh score low. For example, when the decimal point is not recognized even though the segment display is configured to display the decimal point, the calculation unit 16 calculates the seventh score low.

FIG. 11A is an example of an image in which the decimal point is recognized. FIG. 11B is an example of an image in which the decimal point is not recognized. For example, when the display of the segment display is pre-registered to include a decimal point, the 7th score based on the image of FIG. 11A is calculated high, and the 7th score based on the image of FIG. 11B is calculated low. Will be done.

The calculation unit 16 calculates the accuracy of reading the numbers by the reading unit 15 using the first score to the seventh score described above. For example, the calculation unit 16 uses the sum of the first score to the seventh score as the accuracy. The calculation unit 16 may multiply at least a part of the first score to the seventh score by weights, and use the total of them as the accuracy. Alternatively, the calculation unit 16 may use the lowest score of the value calculated using a part of the first score to the seventh score and each of the other parts of the first score to the seventh score as the accuracy. ..

When a plurality of character images are cut out from the input image, the first score to the fourth score are calculated for each character image.
For example, the calculation unit 16 compares a plurality of first scores and uses the lowest first score to calculate the accuracy. Similarly, the calculation unit 16 determines the accuracy of the lowest second score among the plurality of second scores, the lowest third score among the plurality of third scores, and the lowest fourth score among the plurality of fourth scores. Used to calculate.
Alternatively, the calculation unit 16 may calculate the total of the first score to the fourth score for each character image. The calculation unit 16 selects the lowest total value from a plurality of total values, and uses the selected total value to calculate the accuracy. When calculating the total value, the calculation unit 16 may weight each score.

The output unit 17 outputs, for example, information based on the read numbers and accuracy to an external output device. For example, the information includes the numbers read. The information may include results calculated based on the numbers read. The information may include a value calculated based on a plurality of read numbers. The output unit 17 may further output an image in which the reading unit 15 reads the numbers, a time when the numbers are read, and the like. The output unit 17 may output a file containing these data in a predetermined format such as CSV. The output unit 17 may transmit data to an external server using FTP (File Transfer Protocol) or the like. Alternatively, the output unit 17 may perform database communication and insert data into an external database server using ODBC (Open Database Connectivity) or the like.

The processing device 10 includes, for example, a processing circuit including a central processing unit. The functions of the reception unit 11, the preprocessing unit 12, the extraction unit 13, the thinning unit 14, the reading unit 15, the calculation unit 16, and the output unit 17 may be realized by the plurality of processing devices 10. The storage device 20 includes, for example, at least one of a hard disk drive (HDD), a network-attached hard disk (NAS), an embedded multimedia card (eMMC), a solid state drive (SSD), and a solid state hybrid drive (SSHD). .. The processing device 10 and the storage device 20 are connected by wire or wirelessly. Alternatively, the processing device 10 and the storage device 20 may be connected to each other via a network.

FIG. 12 is a flowchart illustrating processing by the reading system according to the embodiment.
The reception unit 11 receives the input image transmitted to the processing device 10 (step St11). The binarization unit 12a binarizes the input image (step St12a). The expansion processing unit 12b expands the binarized input image (step St12b). The shrinking processing unit 12c shrinks the input image that has been binarized and expanded (step St12c). The labeling processing unit 13a assigns a label to a set of white pixels included in the input image (step St13a). The cutout portion 13b cuts out a character image from the input image (step St13b). The thinning unit 14 thins a set of white pixels in an input image or a character image (step St14). The reading unit 15 reads the numbers shown in one character image (step St15). The calculation unit 16 executes the first to fourth processes on one character image, and calculates the first score to the fourth score (step St16a).

Steps St15 and St16a are repeated until i is equal to the number of labels. That is, first, i is set to 0. For one set of white pixels, 1 is added to i when the reading of numbers and the first to fourth processes are completed. Steps St15 and St16a are repeated as long as i is less than the number of labels. That is, when there are a plurality of cut out character images, steps St15 and St16a are executed for each character image.

The calculation unit 16 executes the fifth to seventh processes and calculates the fifth score to the seventh score (step S16b), and the calculation unit 16 calculates the accuracy using the first score to the seventh score. (Step St16c).

The order of the steps shown in the flowchart of FIG. 12 can be changed as appropriate. The fifth to seventh processes may be executed before the first to fourth processes. For example, the reading of step St15 may be executed in parallel with the first to seventh processes. The fifth to seventh processes may be executed in parallel with the first to fourth processes. Alternatively, before the thinning in step St14, a process using the non-thinned character image (for example, a second process, a third process, a fifth process to a seventh process, etc.) may be executed.

The effect of the embodiment will be described.
According to the reading system 1 according to the embodiment, the numbers are read and the accuracy of the reading is calculated. For example, the accuracy allows the user to see how reliable the numbers read are. Even when an abnormal number is output, the user can check whether the number is actually displayed on the segment display based on the accuracy.
The reading system 1 may output an image in which the numbers are read, in addition to the numbers and accuracy. This allows the user to easily see the actually displayed numbers from the image, even when the accuracy is low.
According to the embodiment, the reliability of the reading system 1 can be improved.

In the above-described embodiment, an example is shown in which all of the first to seventh processes are executed and the accuracy is calculated using the first score to the seventh score. The reading system 1 may execute at least one of the first to seventh processes and calculate the accuracy using at least one of the calculated first to seventh scores.

For example, the reading system 1 executes at least one of the first process to the fourth process. The first to fourth processes are processes corresponding to the characteristics of the segment display, and the reliability of accuracy can be improved by executing at least one of these processes. Preferably, the reading system 1 performs at least one of the first and second processes. The first process and the second process are processes corresponding to the display mode of the segment display, and the reliability of the accuracy can be further improved by executing at least one of these processes.

For example, the reading system 1 executes at least one of the first process and the second process, and at least one of the third process and the fourth process. Preferably, the reading system 1 further performs at least one of the fifth to seventh processes. By executing more processes and calculating the accuracy using the scores calculated by those processes, the reliability of the accuracy can be improved.

In the reading system 1, when the characters displayed on the segment display are read, the line thinning process is performed. In the thinning process, a set of pixels indicating characters is thinned. As a result, the width of the set of pixels indicating the characters can be made uniform regardless of the thickness of the characters on the segment display in the input image. Therefore, the characters can be read with high accuracy regardless of the thickness of the characters on the segment display.

Further, in the first process and the fourth process, a plurality of determination areas are set in the character image. Then, in each of the plurality of determination areas, the score is calculated based on the detection result of the number of pixels. By using the detection result of the number of pixels, the score can be calculated regardless of the typeface of the character.

(Example)
An example of the calculation method of the first score to the seventh score in the first process to the seventh process will be described.
In the first process, the calculation unit 16 searches for a straight line along a predetermined direction in each determination area. The initial value of the first score is 0. When a plurality of straight lines are found in the determination area, the first score is added by 5. When a plurality of straight lines are discrete within a predetermined distance, a first score of 5 is added. When the angle exceeds a preset threshold, the first score is added by 2. If no straight line is found, add 5 to the first score. The value obtained by dividing this total value by the pixels × 5 arranged in a predetermined direction is used as the score of the determination area. This process is executed for each determination area. Using the sum of these scores, the value obtained by (7-sum) / 7 is set as the first score.
In the second process, the calculation unit 16 determines the calculated precision as the second score.
In the third process, the calculation unit 16 calculates the ratio R1 of the length of the character image in the first direction D1 and the length of the character image in the second direction D2. The calculation unit 16 refers to the preset ratio R2. The calculation unit 16 calculates V = R1 / R2 when the ratio R2 is larger than the ratio R1. The calculation unit 16 calculates V = R2 / R1 when the ratio R1 is larger than the ratio R2.
The calculation unit 16 uses a value obtained by subtracting the absolute value of the value V from 1.0 as the third score.

In the fourth process, the calculation unit 16 sets the fourth score to "1" if the combination of the determination results matches any of the registered types. If the combination of determination results does not match any of the registered types, the calculation unit 16 sets the fourth score to "0".
In the fifth process, the fifth score is initially set to 1.0. The calculation unit 16 generates a straight line based on the positions of the centers of gravity of a plurality of characters. The calculation unit 16 deducts 0.4 points from the fifth score when the slope of the straight line with respect to the predetermined direction exceeds a preset threshold value. When the distance between the straight line and the position of the center of gravity of the character exceeds a preset threshold value, the fifth score is deducted by 0.2 points. The distance between the straight line and the position of the center of gravity of the character is determined for each character.
In the sixth process, the calculation unit 16 sets the sixth score to "1" when the number of characters is equal to or greater than the minimum number of characters and equal to or less than the maximum number of characters. When the number of characters is less than the minimum number of characters or exceeds the maximum number of characters, the calculation unit 16 sets the sixth score to "0".
In the seventh process, if the result of recognizing whether or not there is a decimal point based on the image matches the information registered in advance, the calculation unit 16 sets the seventh score to "1". The calculation unit 16 sets the seventh score to "0" when the recognition result of the decimal point does not match the information registered in advance.

The calculation unit 16 executes the following calculation.
S ₀ = W ₁ x S ₁ + W ₂ x S ₂ + W ₃ x S ₃ + W ₄ x S ₄
S ₁ is the first score. W ₁ is a weight for the first score. S ₂ is the second score. W ₂ is a weight for the second score. S ₃ is the third score. W ₃ is a weight for the third score. S ₅ is the fifth score. W ₅ is a weight for the fifth score.

The calculation unit 16 _{extracts the lowest score from the calculated scores S 0} , the fourth score, the sixth score, and the seventh score, and uses that score as the accuracy. Therefore, if the 4th score, the 6th score, or the 7th score is 0, the accuracy is 0.

FIG. 13 is a graph showing the reading result and the accuracy in the embodiment.
In FIG. 13, the dark plot indicates whether the reading result by the reading system 1 is correct. When the reading result matches the number shown in the actual image, the value on the vertical axis is "1". Light-colored plots represent the calculated accuracy. The straight line SL represents the average value of accuracy. FIG. 13 shows the result when reading is executed about 140 times.

In the example of FIG. 13, erroneous numbers are read in the 8th, 38th, and 52nd readings. At each of these readings, accuracy was calculated to be significantly lower than the mean. For example, it can be seen from the accuracy that the user should pay attention to the reading results at the time of these readings. In addition, since the accuracy of other reading results is high, it can be seen that even if an abnormal number is read, it is highly possible that the number is actually displayed.

Further, the calculation unit 16 may compare the accuracy with the first threshold value. For example, when the accuracy is lower than the first threshold value, the reading system 1 re-executes the image processing executed when reading the numbers under different conditions. For example, by performing image cropping, distortion correction, expansion processing, contraction processing, and the like under different conditions, it is possible that the numbers can be read accurately.

The reading system 1 may output the read numbers when the accuracy exceeds the first threshold value, and may output the read numbers, the accuracy, and the character image when the accuracy is equal to or less than the first threshold value. The reading system 1 may further output the accuracy when the accuracy exceeds the first threshold value. According to this method, the user only needs to check the character image when the accuracy is low. The time and effort of confirmation by the user can be reduced.

The first threshold is set in advance by the user, for example. Alternatively, the first threshold value may be calculated based on the calculation result of the certainty in the past specified number of times or specified period.

In the above, an example in which the input image is processed by the reading system 1 has been described. Not limited to this example, the processed image may be input to the reading system 1 according to the embodiment. For example, when an image is acquired by an imaging device, a character image is extracted by another processing device. Another processing device transmits the extracted character image to the reading system 1. The reading system 1 may read numbers and calculate accuracy from the received character image.

Further, in the embodiment described above, an example of reading the numbers displayed on the 7-segment display by the reading system 1 has been described. The reading system 1 can be applied to other than the 7-segment display. For example, with respect to the 14-segment display or the 16-segment display, characters can be read by the same processing as described above.

14 (a) and 14 (c) are schematic views illustrating a segment display. 14 (b) and 14 (d) are schematic views illustrating the setting of the determination region.
FIG. 14A shows an example of a 14-segment display. On a 14-segment display, one character is displayed by 14 segments. In this case, the calculation unit 16 sets 14 or more determination areas for the character image. The calculation unit 16 sets the determination area DR of 14 for the character image B, for example, as shown in FIG. 14 (b). The calculation unit 16 executes the first process and the fourth process based on the set determination area DR. Further, the calculation unit 16 executes the second process by using the mask corresponding to the state in which all the segments 14 are lit.

FIG. 14C shows an example of a 16-segment display. On a 16-segment display, one character is displayed by 16 segments. In this case, the calculation unit 16 sets 16 or more determination areas for the character image. The calculation unit 16 sets 16 determination areas DR for the character image B, for example, as shown in FIG. 14D. The calculation unit 16 executes the first process and the fourth process based on the set determination area DR. Further, the calculation unit 16 executes the second process by using the mask corresponding to the state in which all the segments of 16 are lit.

FIG. 15 is a block diagram showing the configuration of the reading system according to the embodiment.
The reading system 1 according to the embodiment includes a processing device 10 and an imaging device 30. In the example of FIG. 15, the reading system 1 further includes a storage device 20 and an output device 40.

The image pickup apparatus 30 photographs the segment display and generates an image. The image pickup apparatus 30 transmits the generated image to the processing apparatus 10. Alternatively, the image pickup device 30 may store the image in the storage device 20. The processing device 10 accesses the storage device 20 and refers to the stored image. When the imaging device 30 acquires a moving image, the imaging device 30 extracts a still image from the moving image and transmits it to the processing device 10. The image pickup apparatus 30 includes, for example, a camera.

The processing device 10 transmits information based on the read characters to the output device 40. The output device 40 outputs the information received from the processing device 10 so that the user can recognize it. The output device 40 includes, for example, at least one of a monitor, a printer, a projector, and a speaker.

The processing device 10, the storage device 20, the image pickup device 30, and the output device 40 are connected to each other by, for example, by wire or wirelessly. Alternatively, they may be connected to each other via a network. Alternatively, at least two or more of the processing device 10, the storage device 20, the image pickup device 30, and the output device 40 may be incorporated in one device. For example, the processing device 10 may be incorporated in an image processing unit or the like of the image pickup device 30.

FIG. 16 is a block diagram showing the configuration of another reading system according to the embodiment.
The reading system 2 shown in FIG. 16 further includes a moving body 50. The moving body 50 moves within a predetermined area. A segment display is provided in the area where the moving body 50 moves. The moving body 50 is, for example, an automatic guided vehicle (AGV). The moving body 50 may be a flying body such as a drone. The moving body 50 may be a robot that walks independently. The moving body 50 may be a forklift, a crane, or the like that performs a predetermined work unmanned.

For example, the processing device 10 and the imaging device 30 are mounted on the moving body 50. The processing device 10 may be provided separately from the mobile body 50 and may be connected to the mobile body 50 via a network. When the moving body 50 moves to a position where the segment display can be photographed, the image pickup apparatus 30 photographs the segment display and generates an image.

As shown in FIG. 16, the reading system 2 may further include an acquisition device 60. The acquisition device 60 is mounted on the moving body 50. For example, an classifier having unique identification information is provided corresponding to the segment display. The acquisition device 60 acquires the identification information possessed by the discriminator.

As shown in FIG. 16, the reading system 2 may further include a control device 70. The control device 70 controls the moving body 50. The moving body 50 moves within a predetermined area based on a command transmitted from the control device 70. The control device 70 may be mounted on the moving body 50 or may be provided separately from the moving body 50. The control device 70 includes, for example, a processing circuit including a central processing unit. One processing circuit may function as both the processing device 10 and the control device 70.

For example, the discriminator is a radio frequency (RF) tag that has ID information. The discriminator emits an electromagnetic field or radio wave containing ID information. The acquisition device 60 receives the electromagnetic field or radio wave emitted from the discriminator and acquires the ID information.

Alternatively, the discriminator may be a one-dimensional or two-dimensional barcode. The acquisition device 60 is a barcode reader. The acquisition device 60 acquires the identification information possessed by the barcode by reading the barcode.

As shown in FIG. 16, the processing device 10 may further include a tying portion 18. For example, when the acquisition device 60 acquires the identification information, it transmits it to the processing device 10. The linking unit 18 associates the transmitted identification information with the read characters and stores them in the storage device 20.

FIG. 17 is a schematic diagram illustrating the operation of another reading system according to the embodiment.
For example, the moving body 50 is a moving body that moves along a predetermined orbit T. The moving body 50 is equipped with an image pickup device 30 and an acquisition device 60. The processing device 10 may be mounted on the moving body 50 or may be provided separately from the moving body 50. The orbit T is provided so that the moving body 50 passes in front of the segment displays SD1 and SD2.

For example, the moving body 50 moves along the orbit T, and when the imaging device 30 reaches a position where the segment display SD1 or SD2 can be photographed, the moving body 50 decelerates or stops. For example, when the moving body 50 decelerates or stops, it transmits an imaging command to the imaging device 30. Alternatively, an imaging command may be transmitted from the control device 70 to the imaging device 30. Upon receiving the command, the image pickup apparatus 30 photographs the segment display SD1 or SD2 while the moving body 50 is decelerating or stopping.

Alternatively, the moving body 50 moves along the orbit T at a speed at which the imaging device 30 can shoot the segment display SD1 or SD2 without blurring. When the imaging device 30 arrives at a position where the segment display SD1 or SD2 can be photographed, an imaging command is transmitted from the moving body 50 or the control device. Upon receiving the command, the image pickup apparatus 30 photographs the segment display SD1 or SD2. When the image pickup device 30 takes a picture and generates an image, the image pickup device 30 transmits the image to a processing device 10 mounted on the moving body 50 or provided separately from the moving body 50.

An identification body ID1 is provided in the vicinity of the segment display SD1. An identification body ID 2 is provided in the vicinity of the segment display SD2. For example, the acquisition device 60 acquires the identification information of the identification body ID1 or ID2 while the moving body 50 is decelerating or stopping.

For example, the moving body 50 moves in front of the segment display SD1. The image pickup apparatus 30 photographs the segment display SD1 and generates an image. The processing device 10 identifies the characters displayed on the segment display SD1 from this image. Further, the acquisition device 60 acquires the identification information of the identification body ID1 corresponding to the segment display SD1. The processing device 10 associates the identified character with the identification information.

Alternatively, when the moving body 50 moves in front of the segment display, the acquisition device 60 may first acquire the identification information of the identifying body. For example, a command is transmitted to the mobile body 50 to read characters on a specific segment display. This directive contains information indicating the position of the segment display. In the storage device 20, each identification information and the position of each identification body are associated and stored. When the acquisition device 60 acquires the identification information, the mobile body 50 accesses the storage device 20. The mobile body 50 refers to the position of the discriminator associated with the identification information. The mobile body 50 determines whether the position of the referenced discriminator matches the position of the segment display instructed to read the characters. If they match, the moving body 50 takes a picture of the segment display SD1 by the image pickup apparatus 30. That is, in this method, the reading of the identification information functions as an interlock when reading the characters on the segment display.

18 and 19 are flowcharts illustrating the processing by the reading system according to the embodiment.
An example of processing by the reading system according to the embodiment will be described with reference to FIGS. 18 and 19.
As shown in FIG. 18, the image pickup apparatus 30 photographs the segment display and acquires an image (first image) (step St21). The acquired image is input to the processing device 10. The processing device 10 reads the numbers on the segment display captured in the image and calculates the reading accuracy (step St22). The processing device 10 determines whether the accuracy exceeds the first threshold value (step St23). When the accuracy is equal to or less than the first threshold value, the shooting conditions are changed (step St24). After that, in step St21, the segment display is photographed again under the changed imaging conditions, and another image (second image) is acquired. When the accuracy exceeds the first threshold value, the processing device 10 adopts the reading result (step S25).

In step St24, for example, at least one of the exposure time, focal length, aperture, and sensitivity of the image pickup apparatus 30 is changed from the shooting conditions when the immediately preceding step St21 is executed. Alternatively, when the reading system includes the moving body 50, the moving body 50 may move in step St24. That is, the position of the moving body 50 may be included as a shooting condition.

Alternatively, when the accuracy is equal to or less than the first threshold value, the imaging device 30 may re-photograph the segment display without changing the imaging conditions. That is, step St24 may be omitted. The segment display is generally lit at a predetermined cycle. Depending on the relationship between the lighting cycle, the shooting timing, and the exposure time, the numbers may be shot unclearly. By shooting the segment display again under the same shooting conditions, a clearer image may be obtained. Further, by not changing the shooting conditions, it is possible to reduce the possibility that the image shot again is too bright or too dark.

Alternatively, the image pickup apparatus 30 may change the shooting conditions according to the results of the processes in the first to seventh processes. For example, when the first or fifth score is low, the image may be too bright or too dark. Therefore, when the first score or the fifth score is low, the image pickup apparatus 30 changes at least one of the exposure time, the aperture, and the sensitivity. When the second score is low, the image may be too bright. When the second score is low, the imaging device 30 performs at least one of shortening the exposure time, increasing the aperture value, and decreasing the sensitivity. For example, in the third process, the calculation unit 16 calculates the ratio of the length of the character image in the first direction D1 to the length of the character image in the second direction D2. Further, the calculation unit 16 calculates the area of the character image. When the proportion is greater than the preset ratio (ratio) and the area is greater than the preset threshold, the image may be too bright. Therefore, the image pickup apparatus 30 executes at least one of shortening the exposure time, increasing the aperture value, and lowering the sensitivity. When the proportion is less than the preset proportion and the area is less than the threshold, the image may be too dark. Therefore, the image pickup apparatus 30 executes at least one of extension of exposure time, reduction of aperture value, and increase of sensitivity. By changing the shooting conditions according to the processing results in the first to seventh processes, the possibility of acquiring a more appropriate image is increased.

In the flowchart shown in FIG. 19, when the accuracy is determined to be equal to or less than the first threshold value in step St23, the processing device 10 determines whether the number of readings executed up to that point exceeds the specified value (step St26). The number of readings is, in other words, the number of executions of step St22. When the number of readings is equal to or less than the specified value, step St24 is executed. When the number of readings exceeds the specified value, the reading system outputs the result without further acquiring the image (step St27). For example, the processing device 10 outputs the read numbers, accuracy, and images. As a result, it is possible to prevent the acquisition of the image from being repeated unnecessarily.

According to the embodiment described above, the reliability of the reading system, the reading method, or the moving body can be improved. By using the program according to the embodiment for operating the computer (processing device) as the reception unit 11, the preprocessing unit 12, the extraction unit 13, the thinning unit 14, the reading unit 15, the calculation unit 16, and the output unit 17. , The reliability of reading can be improved.

The processing of the various data described above is executed based on, for example, a program (software). For example, a computer stores this program and reads this program to process the various information described above.

The above-mentioned processing of various information can be executed by a computer as a program such as a magnetic disk (flexible disk, hard disk, etc.), an optical disk (CD-ROM, CD-R, CD-RW, DVD-ROM, DVD ± R). , DVD ± RW, etc.), semiconductor memory, or other recording medium.

For example, the information recorded on the recording medium can be read by a computer (or an embedded system). In the recording medium, the recording format (storage format) is arbitrary. For example, the computer reads a program from the recording medium and causes the CPU to execute the instructions described in the program based on the program. In the computer, the acquisition (or reading) of the program may be performed through the network.

The processing device and control device according to the embodiment includes one or more devices (for example, a personal computer). The processing device and the control device according to the embodiment may include a plurality of devices connected by a network.

Although some embodiments of the present invention have been illustrated above, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other embodiments, and various omissions, replacements, changes, etc. can be made without departing from the gist of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are also included in the scope of the invention described in the claims and the equivalent scope thereof. Moreover, each of the above-described embodiments can be implemented in combination with each other.

1, 2, 3 reading system, 10 processing device, 11 reception part, 12 pre-processing part, 12a binarization part, 12b expansion processing part, 12c shrinking processing part, 13 extraction part, 13a labeling processing part, 13b cutting part , 14 Thinning unit, 15 Reading unit, 16 Calculator, 17 Output unit, 18 Linking unit, 20 Storage device, 30 Imaging device, 40 Output device, 50 Moving device, 60 Acquisition device, 70 Control device, A Input image, B, B1 to B7 character image, C5 to C7 center of gravity position, CL straight line, D1 first direction, D2 second direction, DR, DR1 to DR7 judgment area, ID1, ID2 discriminator, Li1 to Li4 line, P pixel, SD1 , SD2 segment display, SL straight line, T orbit

Claims (30)

A reading unit that reads the characters from a character image of the characters displayed by the segment display including a plurality of segments.
A calculation unit that executes the first process of calculating the first score and calculates the reading accuracy using the first score.
With
In the first process, the calculation unit
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
In each of the plurality of determination areas, the first score is calculated based on the number of lines included in the thinned pixels indicating the characters and the angle of the lines overlapping the determination areas.
Reading system. The calculation unit
A second process of extracting pixels matching a preset mask from the pixels indicating the characters and calculating a second score based on the matching ratio between the pixels indicating the characters and the extracted pixels.
A third process of calculating a third score based on the ratio of the length of the character image in the first direction to the length of the character image in the second direction intersecting the first direction.
A fourth process of detecting the presence or absence of a part of the character in each of the plurality of determination regions and calculating a fourth score based on a comparison result between the detected result and a preset type.
The reading according to claim 1, wherein at least one of the above is further executed , and the accuracy is calculated by using at least one of the second score, the third score, and the fourth score, and the first score. system. When a plurality of the characters are displayed on the segment display, the reading unit reads the plurality of characters.
The calculation unit
The first process is executed for each of the plurality of character images on which the plurality of characters are copied.
The accuracy is calculated using a plurality of the first scores.
The reading system according to claim 1. A reading unit that reads the characters from a character image of the characters displayed by the segment display including a plurality of segments.
A calculation unit that executes the first process for calculating the first score and the fourth process for calculating the fourth score, and calculates the reading accuracy using the first score and the fourth score.
With
In the fourth process, the calculation unit
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
The presence or absence of a part of the character is detected in each of the plurality of determination areas, and the presence or absence of a part of the character is detected.
The fourth score is calculated based on the comparison result between the detected result and the preset type.
A part of the plurality of determination regions extends in the first direction and is arranged along the first direction, and another part of the plurality of determination regions intersects the first direction. Extending in the direction, lining up along the second direction,
The position of the plurality of determination regions in the first direction, the position in the second direction, the length in the first direction, and the length in the second direction are the length of the character image in the first direction. It is set based on the length of the character image in the second direction .
In the first process, the calculation unit is based on the number of lines included in the thinned pixels indicating the characters in each of the plurality of determination areas and the angle of the lines overlapping the determination areas. A reading system that calculates the first score. The calculation unit
A second process of extracting a pixel matching a preset mask from the pixel indicating the character and calculating a second score based on the matching ratio between the pixel indicating the character and the extracted pixel.
A third process of calculating a third score based on the ratio of the length of the character image in the first direction and the length of the character image in the second direction.
Further performing at least one, at least one of the previous SL second score and the third score, said first score, using a fourth score, according to claim 4, wherein for calculating the accuracy Reading system. When a plurality of the characters are displayed on the segment display, the reading unit reads the plurality of characters.
The calculation unit
The fourth process is executed for each of the plurality of character images on which the plurality of characters are copied.
The accuracy is calculated using the plurality of the fourth scores.
The reading system according to claim 4. The calculation unit further
The fifth process of calculating the fifth score based on the variation in the positions of the plurality of characters and the arrangement direction of the plurality of characters.
The sixth process of calculating the sixth score based on the comparison result of the number of the plurality of characters and the preset minimum number of characters and maximum number of characters.
The seventh process of calculating the seventh score based on the comparison result of the recognition result of the decimal point for the plurality of characters and the presence / absence of the decimal point registered in advance.
The reading system according to claim 3 or 6, wherein at least one of the above is executed, and at least one of the fifth score, the sixth score, and the seventh score is further used to calculate the accuracy. The calculation unit further executes at least the second process,
The reading system according to claim 2 or 5, wherein the mask used in the second process corresponds to a state in which the plurality of segments are lit. The reading system according to any one of claims 1 to 8 , further comprising an output unit for outputting the read character, the accuracy, and the character image. An imaging device that photographs the segment display and acquires a first image including the character image.
An extraction unit that extracts the character image from the first image,
The reading system according to any one of claims 1 to 9, further comprising. The reading system according to claim 10 , wherein the imaging device captures a moving image and cuts out the first image in which the segment display is captured from the moving image. When the accuracy is equal to or less than the first threshold value, the imaging device re-photographs the segment display to acquire a second image.
The reading system according to claim 10 or 11 , wherein the reading unit and the calculation unit perform reading of the characters and calculation of the accuracy of the reading of the second image. The reading system according to claim 12 , wherein the imaging device acquires the second image under the same shooting conditions as when the first image was acquired. The reading system according to any one of claims 10 to 13 , further comprising a mobile body on which the image pickup apparatus is mounted. While reading the character from the character image of the character displayed by the segment display including a plurality of segments,
A reading method in which the first process of calculating the first score is executed and the accuracy of reading is calculated using the first score.
In the first process,
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
In each of the plurality of determination areas, the first score is calculated based on the number of lines included in the thinned pixels indicating the characters and the angle of the lines overlapping the determination areas.
Reading method. Read the plurality of the characters displayed on the segment display,
The first process is executed for each of the plurality of character images on which the plurality of characters are copied.
The accuracy is calculated using a plurality of the first scores.
The reading method according to claim 15. While reading the character from the character image of the character displayed by the segment display including a plurality of segments,
A reading method in which a first process for calculating a first score and a fourth process for calculating a fourth score are executed, and the accuracy of reading is calculated using the first score and the fourth score.
In the fourth process,
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
The presence or absence of a part of the character is detected in each of the plurality of determination areas, and the presence or absence of a part of the character is detected.
The fourth score is calculated based on the comparison result between the detected result and the preset type.
A part of the plurality of determination regions extends in the first direction and is arranged along the first direction, and another part of the plurality of determination regions intersects the first direction. Extending in the direction, lining up along the second direction,
The position of the plurality of determination regions in the first direction, the position in the second direction, the length in the first direction, and the length in the second direction are the length of the character image in the first direction. It is set based on the length of the character image in the second direction .
In the first process, the first score is based on the number of lines included in the thinned pixels indicating the characters in each of the plurality of determination regions and the angle of the lines overlapping the determination regions. Reading method to calculate. Read the plurality of the characters displayed on the segment display,
The fourth process is executed for each of the plurality of character images on which the plurality of characters are copied.
The accuracy is calculated using the plurality of the fourth scores.
17. The reading method according to claim 17. The fifth process of calculating the fifth score based on the variation in the positions of the plurality of characters and the arrangement direction of the plurality of characters.
The sixth process of calculating the sixth score based on the comparison result of the number of the plurality of characters and the preset minimum number of characters and maximum number of characters.
The seventh process of calculating the seventh score based on the comparison result of the recognition result of the decimal point for the plurality of characters and the presence / absence of the decimal point registered in advance.
Do at least one of the more,
The reading method according to claim 16 or 18 , wherein the accuracy is calculated by further using at least one of the fifth score, the sixth score, and the seventh score. Take a video and
The segment display is shown, and the first image including the character image is cut out from the moving image.
The reading method according to any one of claims 15 to 19 , wherein the character image is extracted from the first image. For processing equipment
The characters are read from the character image of the characters displayed by the segment display including a plurality of segments.
The first process of calculating the first score is executed, and the reading accuracy is calculated using the first score.
It ’s a program
In the first process,
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
In each of the plurality of determination areas, the first score is calculated based on the number of lines included in the thinned pixels indicating the characters and the angle of the lines overlapping the determination areas.
program. In the processing device
Have the segment display read a plurality of the characters displayed on the segment display.
The first process is executed for each of the plurality of character images on which the plurality of characters are copied.
The accuracy is calculated using a plurality of the first scores.
21. The program of claim 21. For processing equipment
The characters are read from the character image of the characters displayed by the segment display including a plurality of segments.
The first process for calculating the first score and the fourth process for calculating the fourth score are executed, and the reading accuracy is calculated using the first score and the fourth score.
It ’s a program
In the fourth process,
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
The presence or absence of a part of the character is detected in each of the plurality of determination areas.
The fourth score is calculated based on the comparison result between the detected result and the preset type.
A part of the plurality of determination regions extends in the first direction and is arranged along the first direction, and another part of the plurality of determination regions intersects the first direction. Extending in the direction, lining up along the second direction,
The position of the plurality of determination regions in the first direction, the position in the second direction, the length in the first direction, and the length in the second direction are the length of the character image in the first direction. It is set based on the length of the character image in the second direction .
In the first process, the first score is based on the number of lines included in the thinned pixels indicating the characters in each of the plurality of determination regions and the angle of the lines overlapping the determination regions. Program to calculate. In the processing device
Have the segment display read a plurality of the characters displayed on the segment display.
The fourth process is executed for each of the plurality of character images on which the plurality of characters are copied.
The accuracy is calculated using the plurality of the fourth scores.
23. The program of claim 23. In addition to the processing device
The fifth process of calculating the fifth score based on the variation in the positions of the plurality of characters and the arrangement direction of the plurality of characters.
The sixth process of calculating the sixth score based on the comparison result of the number of the plurality of characters and the preset minimum number of characters and maximum number of characters.
The seventh process of calculating the seventh score based on the comparison result of the recognition result of the decimal point for the plurality of characters and the presence / absence of the decimal point registered in advance.
The program according to claim 22 or 24 , wherein at least one of the above-mentioned fifth score, the sixth score, and the seventh score is further used to calculate the accuracy. A storage medium that stores the program according to any one of claims 21 to 25. A mobile body that moves within a predetermined area provided with a segment display containing a plurality of segments.
Imaging device and
A reading unit that reads characters displayed by the segment display from the image of the segment display taken by the imaging device.
An extraction unit that cuts out a character image that is a copy of the character from the image,
A calculation unit that executes the first process of calculating the first score and calculates the reading accuracy using the first score.
With
In the first process, the calculation unit
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
In each of the plurality of determination areas, the first score is calculated based on the number of lines included in the thinned pixels indicating the characters and the angle of the lines overlapping the determination areas.
Mobile body. A mobile body that moves within a predetermined area provided with a segment display containing a plurality of segments.
Imaging device and
A reading unit that reads characters displayed by the segment display from the image of the segment display taken by the imaging device.
An extraction unit that cuts out a character image that is a copy of the character from the image,
A calculation unit that executes the first process for calculating the first score and the fourth process for calculating the fourth score, and calculates the reading accuracy using the first score and the fourth score.
With
In the fourth process, the calculation unit
A plurality of determination areas corresponding to the plurality of segments are set in the character image, and the character image is set.
The presence or absence of a part of the character is detected in each of the plurality of determination areas, and the presence or absence of a part of the character is detected.
The fourth score is calculated based on the comparison result between the detected result and the preset type.
A part of the plurality of determination regions extends in the first direction and is arranged along the first direction, and another part of the plurality of determination regions intersects the first direction. Extending in the direction, lining up along the second direction,
The position of the plurality of determination regions in the first direction, the position in the second direction, the length in the first direction, and the length in the second direction are the length of the character image in the first direction. It is set based on the length of the character image in the second direction .
In the first process, the first score is based on the number of lines included in the thinned pixels indicating the characters in each of the plurality of determination regions and the angle of the lines overlapping the determination regions. A mobile body that calculates. A control device for controlling the moving body and the imaging device is further provided.
The moving body according to claim 27 or 28 , wherein when the moving body moves to a position where the segment display can be photographed by the imaging device, the control device causes the imaging device to shoot the segment display. The moving body according to any one of claims 27 to 29 , wherein the image pickup apparatus captures a moving image and cuts out the image of the segment display from the moving image.

Images